We're not at that point yet with natural gas because a combined cycle turbine is more efficient than a steam turbine.
Yeah, next 50 years you might not see coal/nat gas being replaced by fusion. But you will see fusion displacing chunks of what those powerplants will be powering
There is no chance that early fusion plants will be small enough to justify building them in the same building as a factory. They will start large.
> For example, aluminum requires ~14-17MWh to produce 1 ton
The Hall–Héroult process runs at 950 C, just below the melting point of copper. It is close to twice the temperature of steam entering the turbines. It is not something that can be piped around casually- as a gas it will always be at very high pressure because lowering the pressure cools it down. Molten salt or similar is required to transport that much heat as a liquid. Every pipe glows orange. Any industrial process will effectively be a part of the power plant because of how difficult it is to transport that heat away.
Also NB that the Hall–Héroult process is for creating aluminum from ore, and recycling aluminum is the primary way we make aluminum.
Industrial parks centered around power plants might become a thing in the future, being looked at as essential infrastructure investment.
Heat transport could be seen as an entire sub-industry unto itself, adding efficiency and cost-savings for conglamorates that choose to partner with companies that invest in and build power plants.
The other guy was correct while you are the one who posted the fallacy. If using heat from nuclear sources to drive aluminum production were feasible, people would already be doing it using heat from HTGR reactors rather than waiting for nuclear fusion reactors to be made. The reason it is not feasible is because the heat is an output, not an input. The actual input is electricity, which is what drives the reaction. The 940–980°C temperatures reached during the reaction are from the electricity being converted into heat from resistive losses.
It should be noted that production nuclear fusion reactors would be even more radioactive than nuclear fission reactors in terms of total nuclear waste production by weight. The only reason people think otherwise is that the hypothetical use of helium-3 fuel would avoid it, but getting enough helium-3 fuel to power even a test reactor is effectively an impossibility. There are many things that are hypothetically attainable if all people in the world decide to do it. The permanent elimination of war, crime and poverty are such things. Obtaining helium-3 in the quantity needed for a single reactor is not.
However, the goal of powering the Hall–Héroult process from a nuclear fusion reactor is doable. Just use solar panels. Then it will be powered by the giant fusion reactor we have in the sky. You would want to add batteries to handle energy needs when the sun is not shining or do a grid tie connection and let the grid operator handle the battery needs.
Finally, industrial processes that actually need heat at high temperatures (up to around 950°C if my searches are accurate) as input could be served by HTGR reactors. If they are not already using them, then future fusion reactors will be useless for them, since there is no future in sight where a man made fusion reactor is a cheaper energy source than a man made fission reactor. Honestly, I suspect using solar panels to harness energy from the giant fusion reactor in the sky is a more cost effective solution than the use of any man-made reactor.
Aluminum reduction is electrochemical, not thermochemical. Yes, the pots are hot, but they are kept hot by resistive dissipation as the alumina is electrolysed.
(There is some chemical energy contributed from oxidation of the carbon electrode.)
For a typical consumer in Sweden, the consumer buys most of their electricity when the prices is at its highest point, which is around 4 times of what the energy cost during the cheapest months. Electricity consumption is at the lowest point when prices are at the lowest, which is the same period when solar peak in production. Inversely, consumption is at the highest when solar production is at the lowest.
That “3X” figure assumes a high‐insolation region (CF ~25 %). In Central Europe, where solar CF is only ~12 %, you’d need about 5x the PV capacity to equal a 1 GW coal plant’s annual generation. How does scaling up to 5 GW of PV change the cost comparison vs a coal plant?
https://www.energy.gov/energysaver/solar-water-heaters
I recall hearing that they are 80% efficient while photovoltaics tend to be around 20% efficient.
As a peer post noted (without back it up but seems reasonable):
> Only 20% of our energy needs are supplied by electricity.
It is a fair viewpoint to talk about energy instead of only electricity. For exemple the current EV are build using charcoal (steel and cement for the infrastructure) and parts/final product are moved around continent with oil (ships). Same for solar panels and their underlying steel structure. Same for the road were using those EV, etc… there’s technical solutions for those, but they didn’t prove to be economically competitive yet. So I’ll happily take that 80% efficiency when we need relatively low heat : domestic and commercial AC and water heating. Those are by far the most energy intensive usage in the residential sector when there isn’t an electric vehicle and are most needs in pick time (mornings, evening at winter). We better take that +60%.
The most economical solution for reducing our carbon emissions by 95% is doing these two steps in parallel:
1. Use electricity instead of fossil fuel 2. Generate electricity in carbon free manner
Yes, there are some use cases this doesn't work well at yet: steel & ocean transport are two you listed. But it does cover the 4 biggest sources of carbon emissions: ground transport, heating, electricity generation and agriculture. The big 4 are 95% of our carbon emissions.
The photovoltaic panels have the added bonus that the energy can be used for other purposes (e.g. transport, HVAC, computers, cooking, laundry, A/V equipment) should my hot water needs be low compared to what the system is designed to produce. However, from a pure efficiency standpoint, it is unclear to me which approach is better. They seem to be a rough tie, with losses for both approaches making the real world worse than ideal conditions. I am not sure if one is better than the other in the actual real world and if anyone who knows the answer is kind enough to share it, I would find the answer enlightening.
That said, in my home, I use net metered photovoltaic panels with a Rheem heat pump for domestic hot water. This was not done because I considered it to be a better solution, but because it was the only solution available to me from local installers.
Solar thermal heating used to make more sense but cost of photovoltaics has come down so much along with relatively cheap heat pump systems nobody is doing the former anymore it seems.
I just got a large solar system installed and next up is a heat pump water heater as thats the second largest user of power next to the HVAC, plus it will cool and dehumidify my garage some where the solar inverter and batteries are located, converting some of the waste heat from the inverter into hot water at the same time.
Batteries beyond the scale of a handheld device only started getting massively manufactured and invested in fairly recently as well. Once it's obvious that it's possible to build megabatteries that can power towns, everyone will want in on the market and prices will go down.
Or inverters? (Also not included in your calc I think?)
Except that it needs to be around 30GW plant to compete with a 1GW coal. And it needs storage for several days of energy.
Fusion would use a conventional turbine with boiling water. Is this a better source of mechanical inertia than hydropower or fission?
Is there a better way to solve the problem of frequency instability?
Why is this fact downvoted? This article mentions "synthetic inertia;" what are its drawbacks?
https://www.bloomberg.com/news/articles/2025-05-09/spain-bla...
Obviously, this configuration of solar and battery banks will work more optimally when they are closer to the equator.
Will different types of power grids be required for areas further away, or is it practical to ship power long distances to far Northern/Southern areas?
Mechanical inertia in generators also tends to do well in these situations.
PV panel supply was just not nearly large enough, and if you look at overall PV capacity as a percentage of their grid capacity, it’s pretty obvious it was never going to be enough to stabilize any serious issues.
Could you point to the outage conclusion report?
You're making the obvious mistake here of equating 1 GW solar with 1 GW of any other source with a 95-99% baseload capacity. To achieve the equivalent result, you'll need to have at least >2 GW actual solar power to equally compare the two.
Granted, in most developed places, solar still beats coal, but this is why in many developing economies with ample coal resources, it makes more sense economically to go with the coal plants.
Take any other resource, say hydel or geothermal - solar and wind quickly go down in economic efficiency terms compared to these, in most cases almost doubling or tripling in costs.
Which is why I compared 1GW of coal power to 3GW of solar power.
We can live with huge land areas converted to power generation, but more space efficient alternatives will be a big improvement.
Conclusion, land isn't really a constraint in the US.
Just pointing out that there are real downsides to this energy source, like all the others.
Now is not the time to stop developing energy sources.
PV provides massively more value per acre than agriculture does. If PV were seriously constrained by land costs, agriculture would be impossible.
But society is perfectly fine with having land producing $500/acre/year of hay, instead of $25,000/acre/year of PV output.
Anyway, the area issue seems not too bad. In the US as least, we have places like the Dakotas which we could turn like 70% of into a solar farm and nobody would really notice.
Since you want to produce power all day, you would take about 20% of that to account for tilt variations and day night cycles, and another 20% to factor in cell efficiency.
So with adequate storage, one square meter of solar can generate an average of 40W of continuous electrical power, 24h per day. Let's round that down to 25W to take into consideration outages and maintenance, bad weather, spacing between panels for personnel access etc.
And there you have it 1GW/25W is about 40 square km with quite generous safety factors, an order of magnitude less than your AI figures. This is still a lot of land if you replace farmland with it, but still totally negligible compared with the millions of square km of hot desert the world has available for this use.
For example, scaling this 400x, to cover for the entire US electrical consumption, is still "only" 16000 sq.km , or 3% of the area of the Great Basins desert in the US, which is one of the smallish deserts of the world compared with Sahara, Ghobi, Kalahari, Australia, Arabia etc. Of course, there is little economic sense to build such a mega-solar farm and pay the cost of energy transport. In practice, we are seeing distributed production taking the cheapest available land nearby.